OK, so your CUDA installation does work, and you're using CUDA
installed through Ubuntu. That latter part is the one difference
between your setup and the one I've been testing with. I can recommend
two things to try:
1) Install your own CUDA to a local directory. Something like:
/bin/bash cuda_11.4.4_470.82.01_linux.run --installpath=<local install path>
You may need to point the script to a new temporary directory if it
complains about not enough space (with --tmpdir=<temp dir name>). Then
set CUDA_HOME to <local install path> and ensure $CUDA_HOME/bin is in
PATH and LD_LIBRARY_PATH contains $CUDA_HOME/lib64. Then try cmake
again.
2) Although the configure script is soon-to-be-deprecated, it may be
worth a shot. With CUDA_HOME properly set try:
./configure -cuda --skip-python gnu
make install
See if that works. If you only want pmemd.cuda you can do 'make pmemd'
instead I think.
Hope this helps,
-Dan
On Thu, Mar 10, 2022 at 6:23 PM Charlotta Lebedenko
<clebedenko.fordham.edu> wrote:
>
> Hi Dan,
>
> I was able to build and run deviceQuery. I had to build as sudo as cuda is
> installed as root. I noticed that CUDART library was statically linked. The
> cmake error for amber install is still about CUDART library missing. Here
> is the test output:
>
> banipsita.MSI:/usr/local/cuda-11.4/samples/1_Utilities/deviceQuery$ sudo
> CUDA_PATH=$CUDA_HOME make
> [sudo] password for banipsita:
> /usr/local/cuda-11.4/bin/nvcc -ccbin g++ -I../../common/inc -m64
> --threads 0 -gencode arch=compute_35,code=sm_35 -gencode
> arch=compute_37,code=sm_37 -gencode arch=compute_50,code=sm_50 -gencode
> arch=compute_52,code=sm_52 -gencode arch=compute_60,code=sm_60 -gencode
> arch=compute_61,code=sm_61 -gencode arch=compute_70,code=sm_70 -gencode
> arch=compute_75,code=sm_75 -gencode arch=compute_80,code=sm_80 -gencode
> arch=compute_86,code=sm_86 -gencode arch=compute_86,code=compute_86 -o
> deviceQuery.o -c deviceQuery.cpp
> nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35',
> 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a
> future release (Use -Wno-deprecated-gpu-targets to suppress warning).
> /usr/local/cuda-11.4/bin/nvcc -ccbin g++ -m64 -gencode
> arch=compute_35,code=sm_35 -gencode arch=compute_37,code=sm_37 -gencode
> arch=compute_50,code=sm_50 -gencode arch=compute_52,code=sm_52 -gencode
> arch=compute_60,code=sm_60 -gencode arch=compute_61,code=sm_61 -gencode
> arch=compute_70,code=sm_70 -gencode arch=compute_75,code=sm_75 -gencode
> arch=compute_80,code=sm_80 -gencode arch=compute_86,code=sm_86 -gencode
> arch=compute_86,code=compute_86 -o deviceQuery deviceQuery.o
> nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35',
> 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a
> future release (Use -Wno-deprecated-gpu-targets to suppress warning).
> mkdir -p ../../bin/x86_64/linux/release
> cp deviceQuery ../../bin/x86_64/linux/release
> banipsita.MSI:/usr/local/cuda-11.4/samples/1_Utilities/deviceQuery$
> ./deviceQuery
> ./deviceQuery Starting...
>
> CUDA Device Query (Runtime API) version (CUDART static linking)
>
> Detected 1 CUDA Capable device(s)
>
> Device 0: "NVIDIA GeForce RTX 3070 Laptop GPU"
> CUDA Driver Version / Runtime Version 11.6 / 11.4
> CUDA Capability Major/Minor version number: 8.6
> Total amount of global memory: 8192 MBytes (8589410304
> bytes)
> (040) Multiprocessors, (128) CUDA Cores/MP: 5120 CUDA Cores
> GPU Max Clock rate: 1620 MHz (1.62 GHz)
> Memory Clock rate: 7001 Mhz
> Memory Bus Width: 256-bit
> L2 Cache Size: 4194304 bytes
> Maximum Texture Dimension Size (x,y,z) 1D=(131072), 2D=(131072,
> 65536), 3D=(16384, 16384, 16384)
> Maximum Layered 1D Texture Size, (num) layers 1D=(32768), 2048 layers
> Maximum Layered 2D Texture Size, (num) layers 2D=(32768, 32768), 2048
> layers
> Total amount of constant memory: 65536 bytes
> Total amount of shared memory per block: 49152 bytes
> Total shared memory per multiprocessor: 102400 bytes
> Total number of registers available per block: 65536
> Warp size: 32
> Maximum number of threads per multiprocessor: 1536
> Maximum number of threads per block: 1024
> Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
> Max dimension size of a grid size (x,y,z): (2147483647, 65535, 65535)
> Maximum memory pitch: 2147483647 bytes
> Texture alignment: 512 bytes
> Concurrent copy and kernel execution: Yes with 5 copy engine(s)
> Run time limit on kernels: Yes
> Integrated GPU sharing Host Memory: No
> Support host page-locked memory mapping: Yes
> Alignment requirement for Surfaces: Yes
> Device has ECC support: Disabled
> Device supports Unified Addressing (UVA): Yes
> Device supports Managed Memory: Yes
> Device supports Compute Preemption: Yes
> Supports Cooperative Kernel Launch: Yes
> Supports MultiDevice Co-op Kernel Launch: No
> Device PCI Domain ID / Bus ID / location ID: 0 / 1 / 0
> Compute Mode:
> < Default (multiple host threads can use ::cudaSetDevice() with device
> simultaneously) >
>
> deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 11.6, CUDA Runtime
> Version = 11.4, NumDevs = 1
> Result = PASS
>
>
> Thank you for your help,
> Charlotta Lebedenko
>
> On Mon, Mar 7, 2022 at 6:38 PM Daniel Roe <daniel.r.roe.gmail.com> wrote:
>
> > On Mon, Mar 7, 2022 at 6:49 AM Daniel Roe <daniel.r.roe.gmail.com> wrote:
> > > @Charlotta, are you able to compile/run any cuda code at all? Maybe the
> > devicequery sample from cuda?
> >
> > Specifically, try the following.
> >
> > Change to the directory <cuda samples>//1_Utilities/deviceQuery
> >
> > Build it (assuming CUDA_HOME is pointing to your CUDA install):
> >
> > CUDA_PATH=$CUDA_HOME make
> >
> > Run deviceQuery
> >
> > You should hopefully get a Result = PASS at the end. If not, something
> > is wrong with the CUDA install.
> >
> > -Dan
> >
> > >
> > > -Dan
> > >
> > >
> > >> CUDA in Windows since it runs
> > >> with a DOS system. If you want to have CUDA support, you can opt to dual
> > >> boot your computer with Ubuntu or any Linux OS. There are installation
> > >> guidelines on the internet.
> > >>
> > >> I hope this helps.
> > >>
> > >>
> > >>
> > >> Cheers,
> > >>
> > >> Alexis
> > >>
> > >> On Mon, Mar 7, 2022 at 2:47 AM Charlotta Lebedenko <
> > clebedenko.fordham.edu>
> > >> wrote:
> > >>
> > >> > Hello,
> > >> >
> > >> > Unfortunately I am still facing this problem with amber installation
> > on
> > >> > windows. I am not sure how to resolve it. Does anyone have any advice
> > about
> > >> > cmake and cuda?
> > >> >
> > >> > Thank you,
> > >> > Charlotta Lebedenko
> > >> >
> > >> >
> > >> >
> > >> > > On Mar 3, 2022, at 15:30, Charlotta Lebedenko <
> > clebedenko.fordham.edu>
> > >> > wrote:
> > >> > >
> > >> > >
> > >> > > Hello,
> > >> > >
> > >> > > Thank you for the advice. We tried these steps and we are still
> > getting
> > >> > the same error. We are doing this on Windows 10, not Windows 11. So
> > maybe
> > >> > there is a problem there.
> > >> > > We set -DCUDA_TOOLKIT_ROOT_DIR and CUDA_HOME to cuda home
> > directory. We
> > >> > probably need help from someone who understands Cmake.
> > >> > >
> > >> > > Here is the error again:
> > >> > >
> > >> > > -- Amber source found, building AmberTools and Amber
> > >> > > -- Could NOT find CUDA (missing: CUDA_CUDART_LIBRARY) (found version
> > >> > "11.4")
> > >> > > CMake Error at cmake/CudaConfig.cmake:15 (message):
> > >> > > You turned on CUDA, but it was not found. Please set the
> > >> > > CUDA_TOOLKIT_ROOT_DIR option to your CUDA install directory.
> > >> > > Call Stack (most recent call first):
> > >> > > CMakeLists.txt:120 (include)
> > >> > >
> > >> > > We also set:
> > >> > >
> > >> >
> > -DCUDA_CUDART_LIBRARY=/usr/local/cuda-11.4/targets/x86_64-linux/lib/libcudart.so.11.4.43
> > >> > >
> > >> > > Our cmake command is:
> > >> > > cmake $AMBER_PREFIX/amber20_src \
> > >> > > -DCMAKE_INSTALL_PREFIX=$AMBER_PREFIX/amber20 \
> > >> > > -DCOMPILER=GNU -DBUILD_QUICK=TRUE \
> > >> > > -DMPI=FALSE -DCUDA=TRUE
> > -DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda-11.4
> > >> >
> > -DCUDA_CUDART_LIBRARY=/usr/local/cuda-11.4/targets/x86_64-linux/lib/libcudart.so.11.4.43
> > >> > -DINSTALL_TESTS=TRUE \
> > >> > > -DDOWNLOAD_MINICONDA=TRUE -DMINICONDA_USE_PY3=TRUE \
> > >> > > 2>&1 | tee cmake.log
> > >> > >
> > >> > > Thank you,
> > >> > > Charlotta Lebedenko
> > >> > >
> > >> > >
> > >> > >
> > >> > >> On Thu, Mar 3, 2022 at 9:32 AM Daniel Roe <daniel.r.roe.gmail.com>
> > >> > wrote:
> > >> > >> Hi,
> > >> > >>
> > >> > >> I've been able to install Amber on Windows11/WSL2 with CUDA with no
> > >> > issue.
> > >> > >> The only different is that I installed CUDA manually (
> > >> > >>
> > >> >
> > https://urldefense.proofpoint.com/v2/url?u=https-3A__developer.nvidia.com_cuda-2D11-2D4-2D4-2Ddownload-2Darchive-3Ftarget-5Fos-3DLinux-26target-5Farch-3Dx86-5F64-26Distribution-3DWSL-2DUbuntu-26target-5Fversion-3D2.0-26target-5Ftype-3Drunfile-5Flocal&d=DwIGaQ&c=aqMfXOEvEJQh2iQMCb7Wy8l0sPnURkcqADc2guUW8IM&r=jOIeI-lTF_2GutlqJUB8jwla6pUZ7k5EeNYDA2RM36A&m=LxDU4uEY3ByKgpRHV5eDbVEPCIZYsLumYeEsQtFAup467V5dapkwob69zo1tel2k&s=iswwSnDzYx_pohoZpVrttYR7-lmM39JF1-6fEZHyMwc&e=
> > >> > )
> > >> > >> into a subdirectory of my home directory and set the environment
> > >> > variables
> > >> > >> CUDA_HOME, PATH, and LD_LIBRARY_PATH, e.g.
> > >> > >>
> > >> > >> export CUDA_HOME=<path_to_cuda>
> > >> > >> export LD_LIBRARY_PATH=<path_to_cuda>/lib64:$LD_LIBRARY_PATH
> > >> > >> export PATH=<path_to_cuda>/bin:$PATH
> > >> > >>
> > >> > >> With those set all I had to do is pass -DCUDA=TRUE to cmake and it
> > was
> > >> > >> picked up just fine.
> > >> > >>
> > >> > >> -Dan
> > >> > >>
> > >> > >> On Wed, Mar 2, 2022 at 4:28 PM Charlotta Lebedenko <
> > >> > clebedenko.fordham.edu>
> > >> > >> wrote:
> > >> > >>
> > >> > >> > Hello,
> > >> > >> >
> > >> > >> > I am still struggling with this same issue during installation.
> > I was
> > >> > >> > hoping someone who has installed Amber with CUDA support on
> > Windows
> > >> > before
> > >> > >> > using WSL2 could offer some advice?
> > >> > >> >
> > >> > >> > Thank you,
> > >> > >> > Charlotta Lebedenko
> > >> > >> >
> > >> > >> > >
> > >> > >> > > On Feb 28, 2022, at 13:04, Charlotta Lebedenko <
> > >> > clebedenko.fordham.edu>
> > >> > >> > wrote:
> > >> > >> > >
> > >> > >> > >
> > >> > >> > > Dear Amber support,
> > >> > >> > >
> > >> > >> > > We are trying to install Amber20 and Ambertools21 with CUDA
> > support
> > >> > on
> > >> > >> > Windows 10 21H2 using WSL2 with ubuntu. We see this cmake error:
> > >> > >> > > -- Could NOT find CUDA (missing: CUDA_CUDART_LIBRARY) (found
> > version
> > >> > >> > "11.4")
> > >> > >> > > CMake Error at cmake/CudaConfig.cmake:15 (message):
> > >> > >> > > You turned on CUDA, but it was not found. Please set the
> > >> > >> > > CUDA_TOOLKIT_ROOT_DIR option to your CUDA install directory.
> > >> > >> > > Call Stack (most recent call first):
> > >> > >> > > CMakeLists.txt:120 (include)
> > >> > >> > >
> > >> > >> > > nvcc is installed at: /usr/local/cuda-11.4/bin/nvcc
> > >> > >> > > CUDA toolkit is installed into a standard location
> > /usr/local/cuda:
> > >> > >> > > ls -l /usr/local
> > >> > >> > > total 36
> > >> > >> > > drwxrwxr-x 2 root root 4096 Feb 28 11:49 bin
> > >> > >> > > lrwxrwxrwx 1 root root 22 Feb 28 11:49 cuda ->
> > >> > /etc/alternatives/cuda
> > >> > >> > > lrwxrwxrwx 1 root root 25 Feb 28 11:49 cuda-11 ->
> > >> > >> > /etc/alternatives/cuda-11
> > >> > >> > > drwxr-xr-x 17 root root 4096 Feb 28 11:49 cuda-11.4
> > >> > >> > >
> > >> > >> > > ls -l /etc/alternatives
> > >> > >> > > lrwxrwxrwx 1 root root 20 Feb 28 11:49 cuda ->
> > /usr/local/cuda-11.4
> > >> > >> > > lrwxrwxrwx 1 root root 20 Feb 28 11:49 cuda-11 ->
> > >> > /usr/local/cuda-11.4
> > >> > >> > >
> > >> > >> > > We set export CUDA_HOME=/usr/local/cuda-11.4
> > >> > >> > >
> > >> > >> > > We added this to run_cmake to point to CUDA toolkit and cudart
> > >> > library:
> > >> > >> > > -DCUDA=TRUE -DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda-11.4
> > >> > >> >
> > >> >
> > -DCUDA_CUDART_LIBRARY=/usr/local/cuda-11.4/targets/x86_64-linux/lib/libcudart.so
> > >> > >> >
> > >> > >> > >
> > >> > >> > > We still see the error above. What else can we try?
> > >> > >> > >
> > >> > >> > > We installed CUDA driver and toolkit following these
> > instructions:
> > >> > >> > > CUDA on WSL :: CUDA Toolkit Documentation (nvidia.com)
> > >> > >> > >
> > >> > >> > > Thank you,
> > >> > >> > > Charlotta Lebedenko
> > >> > >> > _______________________________________________
> > >> > >> > AMBER mailing list
> > >> > >> > AMBER.ambermd.org
> > >> > >> >
> > >> >
> > https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.ambermd.org_mailman_listinfo_amber&d=DwIGaQ&c=aqMfXOEvEJQh2iQMCb7Wy8l0sPnURkcqADc2guUW8IM&r=jOIeI-lTF_2GutlqJUB8jwla6pUZ7k5EeNYDA2RM36A&m=LxDU4uEY3ByKgpRHV5eDbVEPCIZYsLumYeEsQtFAup467V5dapkwob69zo1tel2k&s=Qt2CnTbykcS2KFNj_kieGCGm35YDQhDwppdlYm1Vya0&e=
> > >> > >> >
> > >> > >> _______________________________________________
> > >> > >> AMBER mailing list
> > >> > >> AMBER.ambermd.org
> > >> > >>
> > >> >
> > https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.ambermd.org_mailman_listinfo_amber&d=DwIGaQ&c=aqMfXOEvEJQh2iQMCb7Wy8l0sPnURkcqADc2guUW8IM&r=jOIeI-lTF_2GutlqJUB8jwla6pUZ7k5EeNYDA2RM36A&m=LxDU4uEY3ByKgpRHV5eDbVEPCIZYsLumYeEsQtFAup467V5dapkwob69zo1tel2k&s=Qt2CnTbykcS2KFNj_kieGCGm35YDQhDwppdlYm1Vya0&e=
> > >> > _______________________________________________
> > >> > AMBER mailing list
> > >> > AMBER.ambermd.org
> > >> >
> > https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.ambermd.org_mailman_listinfo_amber&d=DwIGaQ&c=aqMfXOEvEJQh2iQMCb7Wy8l0sPnURkcqADc2guUW8IM&r=jOIeI-lTF_2GutlqJUB8jwla6pUZ7k5EeNYDA2RM36A&m=dDmUDpfZ7phpbAWIypFSiJAtCWKxJ7zE2Q1TrJhbpIjM9dsufD1FExOOQAFtRt-L&s=5YzJQJDkjA4b8YF82kXSTxmQLemTlmCmOzIzts35fPI&e=
> > >> >
> > >> --
> > >> Alexis Azucena
> > >>
> > >> Instructor
> > >>
> > >> School of Technology | University of the Philippines Visayas
> > >> +639175841130
> > >> agazucena.up.edu.ph
> > >> _______________________________________________
> > >> AMBER mailing list
> > >> AMBER.ambermd.org
> > >>
> > https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.ambermd.org_mailman_listinfo_amber&d=DwIGaQ&c=aqMfXOEvEJQh2iQMCb7Wy8l0sPnURkcqADc2guUW8IM&r=jOIeI-lTF_2GutlqJUB8jwla6pUZ7k5EeNYDA2RM36A&m=dDmUDpfZ7phpbAWIypFSiJAtCWKxJ7zE2Q1TrJhbpIjM9dsufD1FExOOQAFtRt-L&s=5YzJQJDkjA4b8YF82kXSTxmQLemTlmCmOzIzts35fPI&e=
> >
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> >
> > https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.ambermd.org_mailman_listinfo_amber&d=DwIGaQ&c=aqMfXOEvEJQh2iQMCb7Wy8l0sPnURkcqADc2guUW8IM&r=jOIeI-lTF_2GutlqJUB8jwla6pUZ7k5EeNYDA2RM36A&m=dDmUDpfZ7phpbAWIypFSiJAtCWKxJ7zE2Q1TrJhbpIjM9dsufD1FExOOQAFtRt-L&s=5YzJQJDkjA4b8YF82kXSTxmQLemTlmCmOzIzts35fPI&e=
> >
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Mar 11 2022 - 12:30:02 PST
